1. Text import
After fifteen minutes with the beta , it transpires to me, that Publisher lacks import or placement of text or table data. Even if I am an idiot and the morning is early, this essential step in digital typesetting should have jumped into my face. Not even pulling a text file from the file manager into the Publisher interface does do anything. If data import is not yet implemented into the beta, it should have been! How are users supposed to test a typesetting program without anything typed?
Being an author and being a typesetter are two different professions, which do not overlap. A typesetter does not copy and paste any text. He does not write anything on his own into the layout, because it is not his name, which will be on the title of the publication. A typesetting program without data import is worthless. 2. XML Workflow
If you publish a new publishing program in 2018, it's all about XML workflow. Nothing else really matters in these days and in the future. The application has to be able to import proprietary XML as well as standardized XML like DOCX, ODT and HTML. A scripting language will be needed, and this can only be Python (not Javascript as in Indesign). 3. File format .afpub
To my horror, the file format .afpub itself turned out to be binary, as if we were living in 1990. It has to be a container with XML! This would have catapulted you ahead of Indesign, which does have all the other XML features apart from the indd file itself. No professional publisher will store layouts in an inaccessible file format.